A pilot study correlating patients' perception of physiological and psychological aspects of fatigue with exercise performance was done to quantify fatigue and its effect on function. Twenty-seven patients with a >6-month history of fatigue and rheumatoid arthritis (RA), polymyositis (PM), or chronic fatigue syndrome (CFS) were evaluated (9 per group). Each completed the following questionnaires: Multidimensional Assessment of Fatigue (MAF); the Human Activity Profile, which provides a score for maximum activity (MAS), adjusted activity (AAS), and a dyspnea scale; the Sickness Impact Profile (SIP), which assesses physical and psychosocial issues; the Profile of Mood States (POMS); and NIH Activity Record (ACTRE). Oxygen consumption and anaerobic threshold were measured during bicycle ergometry to each patient's tolerance. Spearman rank correlation coefficient (rho) was used to rank exercise tolerance (ET) (ET = VO2 measured/VO2 predicted), anaerobic threshold (AT) data, and self-reported measures of fatigue and activity. R values > 0.5 are reported when p <0.02. Patients who reached substantial level of effort and who had low ET also had a low AT (r = 0.788). The MAS was correlated with the measured VO2 (r = 0.56), ET (r = 0.62), the AAS (r = 0.70) and dyspnea (r = 0.57). The AAS was also correlated with the SIP physical mobility scale (r = 0.77). The latter was correlated with dyspnea (r = 0.52) and MAF (r = 0.59). Fatigue reported from the POMS, ACTRE, and MAF did not correlate with either measured VO2 or ET. AT correlated with activity-associated fatigue measured on the ACTRE (r = 0.57). This study showed that physiological measures of fitness did not correlate with self reported measures of fatigue in the aggregated group. Only in patients with CFS did a self-report of fatigue (MAF) correlate with ET (p = 0.05). Seven CFS, 5 RA and 2 PM patients performed below the median of ET, which was 58%. Patients with CFS had the lowest exercise performance of the 3 diagnostic groups. This protocol was terminated in November 1996 at completion of patient recruitment.